AnythingLLM is the ultimate all-in-one desktop AI assistant you have been looking for. AnythingLLM includes a built-in LLM, RAG, AI Agents, and even custom tooling to improve your productivity all while running fully locally & privately on your desktop.
You want a desktop AI that just does it all - that is AnythingLLM.
Hey ProductHunt 👋!
Long-time member, longer-time builder. I wanted to finally formally launch AnythingLLM to the community. AnythingLLM is the all-in-one AI app that just works and does exactly what you hope other apps would do, locally and privately on your desktop.
AnythingLLM bakes in all of the things you are looking for, but with no setup, and no command line tools, all in one application you can just install on your computer.
AnythingLLM includes...
🤖 A built-in LLM engine
That's right, no Ollama, LMStudio, or API key. You can just download the latest cutting-edge models and run them without any other software. You can still use your favorite tool or cloud-based LLM if you have a preference 😁.
📚 Upload and chat with documents
With our built-in embedder and vector database you get full RAG capabilities to chat with your documents like you can in other popular apps. Just drag and drop files or images and you can chat with them nearly instantly.
🧠 AI Agents that are useful
We include some sensible AI agents to help you be productive. This includes tools like Web Search, Web Scraping, SQL Connector, Document Summary, File creation, and soon - custom tools you write or get from our community of over 4K+ members.
🔌 Data connectors
Aside from local documents, you can also connect to various external data sources like GitHub, Confluence, and other websites to build up your LLMs knowledge base.
🔒 Privacy first
Unless you decide to work with a third party, AnythingLLM is fully on-device and private. Everything is stored on your device to ensure maximum security. This is why large corporations opt to use AnythingLLM within their organizations and why you might want to also. Your IT team will love how easy it is to install ❤️
🤩 Open-Source
Lastly, but for the hackers, AnythingLLM is open-source, and with over 18K GitHub stars and 2 years of a solid development track record - AnythingLLM is here to stay.
And some notable other features:
- Text-to-speech and speech-to-text support
- Full developer API for custom integrations
- Poppin Discord community of 4K members
- Dark mode first (and only lol)
- Works with the LLM provider you are already using
AnythingLLM is the result of our team's dedicated focus for 2 years to bring AI fully locally with a no-fuss setup that just works the way you want.
AnythingLLM hopes to be your ideal solution for improving productivity with zero frustrating setups or configurations.
Bonus! We offer a Docker Image with over 100K+ pulls that offers a more "multi-user" and administrative version of AnythingLLM that has all the benefits of the desktop app with rule-based access for those who want a more centralized way to onboarding you org with a cutting-edge AI tool suite.
Let us know what you think!
- Timothy Carambat
I get that you're trying to make an all-in-one desktop AI app but did you consider how most of us already have our preferred AI tools, why should we switch, also does it really run everything locally or are there hidden data transfers going on, trust is a big issue with these things these days isn't it, I mean have you thought about making it more modular, like allowing users to integrate only the AI tools they really need or want and maybe work on the interface a bit, it looks kinda dated
Been using one GPT desktop app but it had limited features. Interesting take with inbuilt agents, will try it out, great job @tim_carambat!
Running an AI assistant on a personal desktop sounds intriguing but with privacy risks of local AI tools and the accuracy of the built-in LLM heavily depending on how updated its training data is, this might not be as seamless as it claims, I'd be cautious about relying entirely on it for productivity without more comprehensive reviews, what do others think.
This new AI tool sounds cool and all but isn't it a bit excessive to have everything crammed into one app, like sometimes I just need simple and specific tools not an entire suite, feels like overkill to me
This is awesome, congrats! Your idea is admirable, I think it is a great idea. How did you come up with this feature?
I use it solely because LanceDB is built into the product. It's totally awesome, and I'm loving it! I can't wait to see what the future holds for this app.
The ability to use it offline and the support for different document types make it very versatile. The seamless integration of LLMs and the focus on local privacy are strong points. I love the idea of managing AI applications and documents all in one place without needing internet access.
This looks like a powerful desktop AI tool. Having everything from custom models to local storage in one place is good for productivity.
The ability to connect to various data sources and use custom tools is a feature that is a feature that many users will find valuable
Finally, an AI assistant that won’t spill my secrets to the cloud! Congratulations on creating such a privacy-focused tool. Love the thoughtful design!
Could you explain more about the custom tools feature? it sounds intriguing and could be a great addition for advanced users
It would love to see more info on how the custom tooling works. Some examples would help visualize the possibilities better.
The download link for the docker image seems to have a minor issue it might need a quick check
Great launch! Running everything locally and privately is such a big plus! It’s refreshing to see a tool focused on privacy without sacrificing performance.
I previously only used cloud LLM services, but AnythingLLM helped me make the jump to trying out all the different flavors of LLMs. Keep up the great work!
The focus on privacy and flexibility with various models and document type is a big plus.
I’m not quite sure how this stacks up against other similar tools. It might be worth comparing its features and performance before diving in.
